Threats and Commitments: International Tribunals and Domestic Trials in Peace Negotiations 
Replication Files
Genevieve Bates

September 5, 2025

***********
README 
***********

These files replicate the creation of the final datasets and the empirical analyses for the paper “Threats and Commitments: International Tribunals and Domestic Trials in Peace Negotiations”. All of the raw data and constituent datasets as well as the code to clean, transform, and merge the data sources were written in R and are included in these replication files. The code to execute the empirical analyses in the main text and the supplementary appendix was written in R and is also included in these replication files.

To replicate, you can either download the full set of files and follow these instructions to create the datasets for analysis or you can download “AnalysisData.RData” and “PanelData.RData” along with “Analysis.Rmd”  without having to reconstruct the analysis datasets from scratch. Both analysis datasets are also available in .csv format as “AnalysisData.csv” and “PanelData.csv” 

*****************************
Replication File Contents
*****************************
The replication files include:

	1.	README.txt
	2.	Codebook.pdf
	3.	Data
		⁃	Raw
			⁃	pax_all_agreements_data.csv
			⁃	V-Dem-CY-Full+Others-v13.rds
			⁃	ISQ2010MurdieDavis_finaltoISQ.dta
			⁃	ICCInvolvementData.csv
			⁃	ged211.csv
			⁃	ucdp-prio-acd-211.csv
			⁃	ICCInvolvementCoding.pdf
			⁃	WDI Data: Note that the World Bank WDI data is not included as a separate file, but the relevant indicators are loaded directly from the "WDI" package in R.
		⁃	Analysis
			⁃	AnalysisData.RData
			⁃	PanelData.RData
			⁃	AnalysisData.csv
			⁃	PanelData.csv
	4.	Code
		⁃	DataClean.Rmd
		⁃	Analysis.Rmd
	5.	Results
		⁃	Figure1.pdf
		⁃	Table1.tex
		⁃	FigureA1.pdf
		⁃	TableA1.tex
		⁃	TableB1.tex
		⁃	TableB2.tex
		⁃	TableB3.tex
		⁃	TableB4.tex
		⁃	TableB5.tex
		⁃	FigureB1a.pdf
		⁃	FigureB1b.pdf
		⁃	FigureB2a.pdf
		⁃	FigureB2b.pdf
		⁃	TableB6.tex
		⁃	TableC1.tex
		⁃	TableC2.tex

The 'Data' folder contains two folders: one folder that includes the raw datasets to be merged to create the final analysis datasets, and one folder the final cross-sectional and panel datasets produced by running “DataClean.Rmd.” Each variable used from these data sources is explained in the Codebook and commented in the “DataClean.Rmd” file. Note: the ‘Raw’ subfolder within the ‘Data’ folder also contains a pdf file, “ICCInvolvementCoding.pdf”, that provides the source information for the original data collection for the ICC involvement dataset. Please also note that the datasets in the ‘Analysis’ folder will be overwritten with identical copies while following the instructions in the “DataClean.Rmd” file.

Results
The results folder contains the .pdf and .tex files for the main tables and figures included in the main text and the supplementary appendices. The .tex files were then manually edited in Overleaf to format for publication.

Codebook
A full codebook, including any transformed data from the raw data sources, is included in these replication files in a document called “Codebook.pdf.”

***************************
Replication Instructions
***************************
To replicate the dataset construction and empirical analyses, download these replication files and run the .Rmd files in order (first running “DataClean.Rmd,” then running “Analysis.Rmd.” The first .Rmd file reads in and merges various data sources to create cross-sectional and panel analysis datasets. The second .Rmd file (“Analysis.Rmd”) executes the analyses and produces all tables and figures presented in the paper and supplementary appendices.

To replicate the construction of the “AnalysisData” and “PanelData” datasets, users should take the following steps:
1. Open the first replication file: DataClean.Rmd
2. On line 16, set the working directory as the folder containing the replication folder downloaded from the APSR Dataverse.
3. Run each chunk of code in order: It begins by reading in and cleaning the VDem data, then loading and cleaning the key variables from the World Bank, then loading and cleaning the human rights organization data, then loading the ICC involvement and ratification data, then loading cleaning and transforming the UCDP Geo-referenced Events and Armed Conflict data. It then merges each dataset into a single panel data frame, transforms relevant variables by logging them, and lags the relevant variables. It then writes and saves the complete panel dataset as .RData and .csv files.

It then loads the PA-X peace agreement data, cleans and transforms the relevant variables, then merges the panel dataset with the cleaned PA-X data to produce the dataset used for the primary analysis. It ends by writing and saving the analysis dataset as .RData and .csv files.


To replicate the empirical analysis and reproduce all tables and figures included in the main text of the article and all supplementary appendices, users should take the following steps:
1. Open the second replication file: Analysis.Rmd
2. On line 17, set the working directory as the folder containing the replication folder downloaded from the APSR Dataverse.
3. Run each chunk of code in order. It produces each table and figure in the order presented in the main text of the article, then produces each table and figure in the order presented in the supplementary appendices. The Markdown file outline and comments throughout the code clearly identify which table or figure is being produced by each chunk of code.

************************************
Data Sources for Analysis Data
************************************
This section explains what variables in the analysis are directly coming from each of the datasets included in the ‘Raw’ data folder. Note that this includes data that was transformed and/or cleaned while running the "DataClean.Rmd" file. For explanations of each variable and their citation please see the Codebook.
	1.	pax_all_agreements_data.csv (Bell et. al 2020)
		⁃	Agreement includes a trial commitment (natcourts)
		-	Agreement includes a trial commitment or a prohibition of amnesty for international crimes (natcourts2)
		- 	Agreement includes a strong trial commitment (natcourts3)
		⁃	Agreement includes any transitional justice commitment (tj)
		⁃	Agreement includes signature from the UN or an IO other than the UN (ic_UNOth)
	2.	V-Dem-CY-Full+Others-v13.rds (Coppedge et al. 2020)
		⁃	participatory democracy (v2x_partipdem)
		⁃	access to justice (v2xcl_acjst)
	3.	World Bank World Development Indicators (Arel-Bundock 2025)
		⁃	GDP per capita (lag_loggdppc)
		⁃	Development assistance as percent of GNI (lag_odagni)
	4.	ISQ2010MurdieDavis_finaltoISQ.dta (Smith & West 2005; Murdie et al. 2012)
		⁃	human rights org presence (hrfilled)
	5.	ICCInvolvementData.csv
		⁃	Rome Statute Ratification (ICC_ratification)
		⁃	ICC institutional threat (threati)
		⁃	ICC intervention (intervention)
		⁃	UN Security Council referral (unscref)
		⁃	Article 14 referral (art14ref)
		⁃	Prop Motu referral (propmo)
		⁃	Stage of ICC involvement (stage)
		⁃	Extent of ICC involvement (extent)
		⁃	ICC minimalist threat (threatm)
	6.	ged211.csv (Sundberg & Melander 2013)
		⁃	casualties (lag_casualties)
		⁃	atrocities (log_atrpost2002)

***************************
Software Dependencies
***************************
All analyses were replicated in R version 4.4.0 (“Puppy Cup”) on the Mac OS Ventura 13.7.6 operating system

* R packages for DataClean.Rmd
	- tidyverse 2.0.0
	- haven 2.5.4
	- dplyr 1.1.4
	- readr 2.1.5
	- readxl 1.4.3
	- zoo 1.8.12
	- countrycode 1.6.0
	- WDI 2.7.8
	- date 1.2.42

* R packages for Analysis.Rmd
	- tidyverse 2.0.0
	- haven 2.5.4
	- dplyr 1.1.4
	- readr 2.1.5
	- zoo 1.8.12
	- date 1.2.42
	- readxl 1.4.3
	- glm2 1.2.1
	- ggplot2 3.5.1
	- brglm 0.7.2
	- brglm2 0.9.2
	- xtable 1.8.4
	- modelsummary 2.0.0
	


